EP0519442A1

EP0519442A1 - CCD image sensor

Info

Publication number: EP0519442A1
Application number: EP92110258A
Authority: EP
Inventors: Naoki Kato
Original assignee: Sony Corp
Current assignee: Sony Corp
Priority date: 1991-06-21
Filing date: 1992-06-17
Publication date: 1992-12-23
Anticipated expiration: 2012-06-17
Also published as: DE69227338T2; EP0519442B1; US5379064A; JPH04373274A; DE69227338D1

Abstract

A CCD imager includes at least one four-phase vertical shift register having first, second, third and fourth transfer electrodes (G₁ to G₄). In order to prevent nonuniform distribution of smear charge and other unwanted charge, and for improving the quality of a picture, means are provided for making the potential well under each first electrode (G₁) deeper than the potential wells under the second, third and fourth electrodes (G₂ to G₄). Preferably, a N-type semiconductor region (6) is formed under each first electrode (G₁) by ion implantation.

Description

BACKGROUND OF THE INVENTION

The present invention relates to a CCD imager having 4-phase vertical shift registers operated by three-valued clock signals (having readout level, high level and low level).
In general, a four-phase CCD solid-state imager has at least one photosensitive imaging section, and one ore more shift registers having four clock input terminals connected with a driver circuit 21 (shown in Fig. 8) by four clock lines. Field readout operations of such a CCD solid-state imager is performed as shown in Figs. 5A-6B. During a readout period, readout pulses (of 15V) are applied, respectively, to first transfer electrodes G₁ for receiving a first drive pulse signal øV₁, and third transfer electrodes G₃ for receiving a third drive pulse signal øV₃. In the odd field, as shown in Figs. 5A and 5B, signal charge packets are transferred from photosensitive elements S_n, S_n+1, S_n+2, S_n+3, ··· to corresponding potential wells which are formed independently under the first and third transfer electrodes G₁ and G₃. Thereafter, the potential barrier under each second transfer electrode G₂, is lowered to create a wide potential well extending continuously under the first, second and third electrodes G₁, G₂ and G₃ of each stage. Thus, one continuous potential well is formed for each of pairs (S_n, S_n+1), (S_n+2, S_n+3), ···. Therefore, the signal charge of the nth photosensitive element Sn of the nth row is mixed with the signal charge of the (n+1)st photosensitive element Sn+1 of the (n+1)st row in the corresponding continuous wide potential well. The next continuous potential well serves for mixing the signal charge packets of the next pair (S_n+2, S_n+3). In this way, the signal charge of each photosensitive element is mixed with the signal charge of the mate of the pair. In these figures, signal charge is shown by parallel oblique lines.
In the even field, the photosensitive elements are paired to form pairs (S_n+1, S_n+2), (S_n+3, S_n+4), ···. First, signal charge packets of the photosensitive elements are transferred to respective independent potential wells formed under the the first and third transfer electrodes G₁ and G₃. Then, lowering the potential barrier of each fourth transfer electrode G₄ creates a continuous wide potential well extending continuously under the third, fourth and first electrodes G₃, G₄ and G₁. Therefore, the continuous wide potential wells function to mix the signal charge packets of the (n+1)st row and (n+2)nd row, the signal charge packets of the (n+3)rd row and (n+4)th row, and so on.
Unwanted charge, such as smear charge and dark signal charge, generated in the vertical shift registers are also transferred in the same manner as explained above for the signal charge. Figs. 7 and 8 show the clock pulse waveforms for the vertical registers and profiles of potential in the odd field. Figs. 9 and 10 shows the clock pulse waveforms and potential profiles in the even field. In these figures, unwanted charge including smear charge and dark signal charge is shown by parallel oblique lines.
In the odd field, the unwanted charge packet stored under the first and second transfer electrodes G₁ and G₂ of a given electrode set (the first electrode set, for example) at an instant t₀ is shifted to the right as viewed in Fig. 8, and this unwanted charge packet exists in the location under the first and second electrodes G₁ and G₂ of the next electrode set (the second set, for example) at an instant t₈, as shown in Fig. 8. At this instant t₈ (t=t₈), the first and second drive pulse signals øV₁ and øV₂ are held at the high level (0 V), and the third and fourth drive pulse signals øV₃ and øV₄ are at the low level (-9 V). At the next instant t₉, the third drive pulse signal øV₃ has switched to the high level, and therefore, the unwanted charge is accumulated under the first, second and third electrodes G₁, G₂ and G₃. At the next instant t₁₀, the second drive pulse signal øV₂ is at the low level, and therefore, the unwanted charge is divided into a first portion under the first electrode G₁ and a second portion under the third electrode G₃.
At an instant t₁₁, the second drive pulse signal øV₂ has switched again to the high level, and the unwanted charge is transfer and accumulated again under the first, second and third transfer electrodes G₁, G₂ and G₃. During the readout period T after the instant t₁₀, a readout pulse p (15 V) is applied to each of the first and third transfer electrodes G₁ and G₃. Therefore, signal charge packets of the sensing elements are transferred to the associated vertical shift register, and accumulated in the respective storage sites under the first, second and third electrodes G₁, G₂ and G₃ at t₁₁. The storage levels of the signal charge are shown by one dot chain lines in Fig. 8.
In the even field, as shown in Figs. 9 and 10, unwanted charge such as smear charge is shifted during an interval between t₀ and t₈, from the location under the first and second transfer electrodes G₁ and G₂ of the first electrode set, for example, to the location under the first and second transfer electrodes G₁ and G₂ of the second electrode set, as in the odd field. Then, the third drive pulse signal øV₃ is switched to the high level, so that, at an instant t₉, the first, second and third drive pulses signals øV₁, øV₂ and øV₃ are high and the fourth drive pulse signal øV₄ is low as shown in Fig. 9. Therefore, the unwanted charge is accumulated at t₉ under the first, second and third transfer electrodes G₁, G₂ and G₃ of each set. At an instant t₁₀, because of the most recent fall of the second drive pulse signal øV₂ to the lower level between t₉ and t₁₀ in Fig. 9, the unwanted charge is divided into a first portion under the first transfer electrode G₁ and a second portion under the third transfer electrode G₃, by a potential barrier formed under each second transfer electrode G₂.
At an instant t₁₁, the fourth drive pulse signal øV₄ has risen to the high level, and the unwanted charge is accumulated under the third, fourth and first transfer electrodes G₃, G₄ and G₁. During the readout period T between t₁₀ and t₁₁, the readout pulses p are applied to the first and third transfer electrode G₁ and G₃ as shown in Fig. 9, so that signal charge packets are taken out from the sensing elements and accumulated at the instant t₁₁ under the third, fourth and first electrodes G₃, G₄ and G₁, as shown by one dot chain lines in Fig. 10.
In the conventional CCD imager, however, nonuniform distribution of unwanted non-signal change in the even field causes noise in reproduced picture imagery. As shown in Fig. 10, the unwanted charge residing in each temporary storage site at t₉ is divided into two portions at t₁₀ due to a potential barrier formed under each second transfer electrode G₂ with the second drive pulse signal øV₂ at the lower level. In this case, the allotment is not uniform but differs in different locations. The ratio between two portions into which an unwanted charge packet is divided is affected by irregularity in pattern and fabricating process. In the example shown in Fig. 10, each unwanted charge packet existing at t₉ is split at t₁₀ in two unequal parts, and the ratio of one part to the other is 4:6 in one packet, and 8:2 in the next packet.
At the instant t₁₁, a potential well is formed under each fourth transfer electrode G₄ by the fourth pulse signal switched to the high level while the potential barrier under each second transfer electrode G₂ still remains. Therefore, the unwanted charge remains nonuniformly distributed among a plurality of wide potential wells of a threefold width (as schematically shown in Fig. 10) formed, respectively, under a plurality of three consecutive electrode sets of the third, forth and first transfer electrodes G₃, G₄ and G₁. The signal charge from the sensing elements is accumulated in the threefold wide potential wells among which the unwanted charges are nonuniformly distributed. As a result, the level of accumulated charge differs from bit to bit in the vertical register. In the example shown in Fig. 10, the amount of accumulation is great in one potential well W₂ while the neighboring potential wells W₁ and W₃ on both sides are filled only to lower levels. Because of this irregularity, the quality of pictures is significantly degraded by unseemly noises (black-and-white point defects) N appearing, in a picture 11 as shown in Fig. 11, in positional relation in the vertical direction with an image of a bright object A.
In the odd field, on the other hand, the uneven distribution of the non-signal charge appearing at the instant t₁₀ in Fig. 8 is only temporary one which disappears shortly and causes no defects in the reproduced picture. At the next instant t₁₁ in Fig. 8, the same potential well is formed as at the instant t₉ under each neighboring group of the first, second and third transfer electrodes G₁, G₂ and G₃. Therefore, the distribution of the unwanted charges is restored to the uniform state existing at the instant t₉.

SUMMARY OF THE INVENTION

It is therefore an object of the present invention to provide a solid-state imager which can prevent nonuniform distribution of charge and improve the quality of reproduced pictures.
According to the present invention, a solid-state imager comprises at least one four-phase vertical shift register. The four-phase vertical shift register comprises at least one first transfer electrode to which a first drive pulse signal is to be applied, at least one second transfer electrode to which a second drive pulse signal is to be applied, at least one third transfer electrode to which a third drive pulse signal is to be applied, and at least one fourth transfer electrode to which a fourth drive pulse signal is to be applied. Each of the transfer electrodes is formed on an insulating layer which is formed on a top surface of a semiconductor substrate. This vertical shift register is so connected with a photosensitive imaging section that photogenerated signal charge packets are transferred from the imaging section to the vertical shift register by applying a readout voltage to the first and third transfer electrodes. In the solid-state imager according to the present invention, a potential well under the first transfer electrode is deeper than a potential well under any of the second, third and fourth transfer electrodes.
This solid-state imager according to the present invention can accumulate unwanted charge such as smear charge in the deeper potential well under the first transfer electrode, and prevent the unwanted charge from moving from under the first transfer electrode to next storage sites under the second and third transfer electrodes. In this way, this solid-state imager can prevent the unwanted charge from being divided into two unequal parts, by temporarily trapping the unwanted charge under the first transfer electrode.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1A is a schematic sectional view showing a main portion (a vertical shift register and its transfer stage) of a CCD solid-state imager according to one embodiment of the present invention.
Fig. 1B is a view showing a potential profile in the vertical shift register shown in Fig. 1A with all the transfer electrodes G₁, G₂, G₃ and G₄ held at the same potential.
Figs. 2A and 2B are cross sectional views showing one process for fabricating the structure shown in Fig. 1A.
Fig. 3 is a schematic view showing operations of the vertical shift register shown in Fig. 1A in the even field.
Figs. 4A and 4B are sectional views showing a process for fabricating a modification of the structure shown in Fig. 1A. Fig. 4C is a view showing a potential profile in the vertical shift register shown in Fig. 4B with the first and second transfer electrodes G₁ and G₂ at a high level potential, and the third and fourth transfer electrodes G₃ and G₄ at a low level potential.
Figs. 5A and 5B are schematic sectional view and schematic plan view of a vertical shift register in the odd field of a field readout mode.
Figs. 6A and 6B are schematic sectional view and schematic plan view of a vertical shift register in the even field of the field readout mode.
Fig. 7 is a time chart showing sequentially clocked vertical drive pulses øV₁, øV₂, øV₃ and øV₄ in the odd field in a conventional example.
Fig. 8 is a schematic view showing operations of a conventional vertical shift register in the odd field.
Fig. 9 is a time chart showing the vertical drive pulses in the even field.
Fig. 10 is a schematic view showing operation of the conventional vertical shift register in the even field.
Fig. 11 is a schematic view showing point defects in a picture produced by the conventional imager.

DETAILED DESCRIPTION OF THE INVENTION

Figs. 1A-4 show one embodiment of the present invention. A CCD imager according to this embodiment has vertical shift registers one of which is shown in Fig. 1A. This example employs a P-type substrate 1. In the section of Fig. 1A, an N-type diffusion layer 2 is formed on the P-type underlying layer of the substrate 1. The imager has a plurality of electrodes sets each of which consists of first, second, third and fourth transfer electrodes G₁, G₂, G₃ and G₄. Each of the transfer electrodes G₁, G₂, G₃ and G₄ is formed on a gate insulating layer 3 which is in turn formed on the top surface of the N-type layer 2. In this example, the second and fourth transfer electrodes G₂ and G₄ are formed by a first (lower) polycrystalline silicon layer, and the first and third transfer electrodes G₁ and G₃ are formed by a second (upper) polycrystalline silicon layer. These transfer electrodes G₁, G₂, G₃ and G₄ are regularly arrayed (as shown in Fig. 3) in the vertical direction (the y direction in Fig. 1A) to form each of the vertical shift registers with the N-type top layer 2.
The first transfer electrodes G₁ are all connected to an first clock input terminal for receiving a first drive pulse signal øV₁ from a driver circuit (or driving means) 21 (shown in Fig. 8). The second transfer electrodes G₂ are all connected to a second clock input terminal for receiving a second drive pulse signal øV₂ of the driver circuit 21. There are further provided a third clock input terminal, connected with all the third transfer electrodes G₃, for receiving a third drive pulse signal øV₃, and a fourth clock input terminal, connected with all the fourth transfer electrodes G₄, for receiving a fourth drive pulse signal øV₄.
In this embodiment, there is further formed, under each first transfer electrode G₁, a region 6 for forming a potential step ph shown in Fig. 1B by making the impurity concentration higher under each first electrode G₁ than under any of the second, third and fourth electrodes G₂, G₃ and G₄. In this embodiment, the region 6 is a highly doped N⁺-type region extending into the N-type top layer 2 from the top surface of the N-type layer 2. When the same voltage is applied to all the electrodes G₁, G₂, G₃ and G₄, the potential is made higher only under the first electrodes G1, as shown in Fig. 1B, than the potential under the second, third and fourth electrodes G₂, G₃ and G₄.
Figs. 2A and 2B show one fabricating process for the structure shown in Fig. 1A. As shown in Fig. 2A, the N-type top layer 2 used as the vertical shift register is first formed on the P-type silicon substrate 1. Then, the gate insulating film 3 of SiO2 is formed, on the top surface of the semiconductor substrate, by thermal oxidation.
Thereafter, as shown in Fig. 2B, the second and fourth electrodes G₂ and G₄ are formed by depositing and patterning the first polysilicon layer. Then, the highly doped N⁺-type regions 6 are formed in the N top layer 2 by a step of ion implantation introducing the N-type impurities such as phosphorus (P) or arsenic (As), using a mask 5 having windows 4 opened at positions at which the first electrodes G₁ are to be formed.
Then, as shown in Fig. 1A, the first and third transfer electrodes G₁ and G₃ are formed by patterning the second polysilicon layer. In this way, the impurity concentration can be made higher only under the first electrodes G₁. The impurity concentration is substantially uniform in the N top layer 2 along the y direction under the second, third and fourth transfer electrodes G₂, G₃ and G₄.
The CCD imager according to this embodiment of the invention is operated as shown in Fig. 3. The vertical shift register shown in Fig. 1A is clocked in the even field by the four drive pulse signals shown in Fig. 9 In Fig. 9, clock timings in an interval between t₀∼t₈ are determined according to an ordinary four-phase clock timing schedule.for line shift (vertical scanning). The period T is a readout period for transferring signal charge from the sensing elements of the photosensitive imaging section to the vertical shift registers.
At an initial instant t₀ shown in Figs. 3 and 9, the first and second drive pulse signals øV₁ and øV₂ are in the high level, and the third and fourth drive pulse signals øV₃ and øV₄ are in the low level. This state is an initial state of the vertical transfer. In this state, a continuous potential well is formed under each neighboring pair of the first and second transfer electrodes G₁ and G₂ held at the high level. In this state, unwanted non-signal charge such as smear charge and dark signal charge is accumulated in the potential step ph formed under each of the first transfer electrodes G₁, as shown in Fig. 3. The unwanted charge is shown by parallel oblique lines in Fig. 3.
At the first point t₁ in time, the third drive pulse signal øV₃ is in the high level together with the first and second drive signals øV₁ and øV₂, and therefore the potential well under each neighboring pair of the first and second electrodes G₁ and G₂ extends under the next third electrode G₃. However, the unwanted charge is still held in the potential step ph under each first electrode G₁.
At the second point t₂ in time, the first drive pulse signal øV₁ is switched to the low level, and lowers the potential under each first electrode G₁. Therefore, the unwanted charge is transferred in the vertical shift direction, and stored under the second and third electrodes G₂ and G₃.
At the instant t₃, the fourth drive signal øV₄ is held in the high level together with the second and third signals øV₂ and øV₃, and the unwanted charge is stored in the potential well formed under each neighboring group of the second, third and fourth electrodes G₂, G₃ and G₄.
At the instant t₄, the second drive pulse signal øV₂ has been switched to the low level, and a potential barrier is formed under each second transfer electrode G₂. Therefore, the unwanted charge is stored under each neighboring pair of the third and fourth electrodes G₃ and G₄.
At the instant t₅, the potential is made higher again under each first electrode G₁ by the first drive pulse signal øV₁ switched to the high level, and therefore, the unwanted charge is collected again in the potential step ph beneath each first electrode G₁.
At the instant t₆, a potential barrier is formed under each third electrode G₃ with the third drive pulse signal øV₃ in the low level. Nevertheless, the unwanted charge is still held in the potential step ph under each first electrode G₁.
At the instant t₇, a potential well is formed under each second electrode G₂ by a rise of the second drive pulse signal øV₂ to the high level. The unwanted charge yet remains in the potential step ph under each first electrode G₁.
At the instant t₈, the fourth drive pulse signal øV₄ is in the low level, and a potential barrier is formed under each fourth electrode G₄. The unwanted charge is still in the potential steps ph.
At the instant t₉, the third drive pulse signal øV₃ switched to the high level forms potential wells under the third electrodes G₃. At the ninth instant t₉, there are formed a plurality of threefold wide stepped potential wells each of which is formed under a neighboring group of the first, second and third transfer electrodes G₁, G₂ and G₃. Each of these stepped wells is deeper only under the first electrode G1, and slightly shallower under the second and third transfer electrodes G₂ and G₃. Therefore, the unwanted charge is still trapped under the first transfer electrode G₁.
At the instant t₁₀, potential barriers are formed under the second electrodes G₂ with the second drive pulse signal øV₂ in the low level. In this case, the unwanted charge in each potential step ph is not divided into two portions, but all of the unwanted charge is persistently confined in each potential step ph. In the case of the conventional device, the formation of a potential barrier appearing under each second electrode at the instant t₁₀ causes uneven distribution of the unwanted charge as shown in Fig. 10. In the vertical shift register according to this embodiment of the invention, by contrast, the potential steps ph prevent nonuniform redistribution of the unwanted charge by holding the unwanted charges under the first electrodes G₁.
At the instant t₁₁, although potential wells are formed under the fourth electrodes G₄ with the fourth signal øV₄ in the high level, the unwanted charge is invariably held in the potential steps ph under the first electrodes G₁. During the readout period T between t₁₀ and t₁₁, signal charge packets are transferred from the sensing elements to the vertical shift register by the readout pulses applied to the first and third electrodes G₁ and G₃, and accumulated in the potential wells under neighboring groups of the third, fourth and first electrodes G₃, G₄ and G₁, as shown by one dot chain lines in Fig. 3.
Thereafter, the signal charge packets are transferred sequentially to a horizontal shift register under the control of the four-phase drive pulses øV₁∼øV₄ of proper phases, together with the unwanted charge.
Fig. 4A shows a variation of the embodiment, in which the highly doped N⁺-type region 6 is formed only partially under each first electrode G₁. In the example shown in Fig. 4A, the highly doped N⁺-type region 6 is formed by ion implantation only under the right half of each first electrode G₁. As shown in Fig. 4A, a mask 5 is aligned to protect the silicon surface under the left halves of the first electrodes G₁ against the implanted ions. The original doping level of the N-type layer 2 is held unchanged under the left halves of the first electrodes as well as under the second, third and fourth electrodes G₂, G₃ and G₄, as shown in Fig. 4B.
When the high level potential is applied to the first and second electrodes G₁ and G₂ and the low level potential is applied to the third and fourth electrodes G₃ and G₄, then a potential step ph is formed only in a portion corresponding to each highly doped region 6, as shown in Fig. 4C. Therefore, this structure can not only accumulate unwanted charge packets but also improve the vertical transfer efficiency of signal charge packets.
In the examples shown in Figs. 1A and 4A, the highly doped regions 6 are used as a means for creating the potential steps ph under the first transfer electrodes G₁. However, it is possible to create potential steps only under the first electrodes G₁ by setting the first drive pulse signal øV₁ at a slightly higher bias than the bias conditions of the other three drive pulse signals øV₂, øV₃ and øV₄.
In this way, the CCD imagers according to the present invention can prevent nonuniform distribution of smear charge, and prevent defects in reproduced pictures.

Claims

A solid-state imager comprising:
a photosensitive imaging section;
first, second, third and fourth input terminals for receiving, respectively, first, second, third and fourth four-phase drive pulse signals; and
a four-phase vertical shift register comprising a first transfer electrode connected with said first input terminal, a second transfer electrode connected with said second input terminal, a third transfer electrode connected with said third input terminal, and a fourth transfer electrode connected with said fourth input terminal, each of said transfer electrodes being formed on an insulating layer which is formed on a top surface of a semiconductor substrate, said vertical shift register being connected with said imaging section so that signal charge packets are transferred from the imaging section to the vertical shift register by applying a readout voltage to each of said first and third transfer electrodes;
wherein a potential well under said first transfer electrode is deeper than a potential well under any of said second, third and fourth transfer electrodes.
A solid-state imager according to Claim 1 wherein said substrate comprises a P-type underlying layer, and an N-type buried channel extending on said underlying layer.
A solid-state imager according to Claim 2 wherein said substrate further comprises an N-type ion implanted region which is formed under said first transfer electrode in said N channel to make the potential well under said first transfer electrode deeper than any of the potential wells under said second, third and fourth transfer electrodes.
A solid-state imager according to Claim 3 wherein said N-type ion implanted region is formed by ion implantation with an N-type dopant of phosphorus.
A solid-state imager according to Claim 3 wherein said N-type ion implanted region is formed by ion implantation with an N-type dopant of arsenic.
A solid-state imager according to Claim 3 wherein each of said first and third drive pulse signals assumes one of three different values.
A CCD imager comprising;
a top layer of a first conductivity type formed in a semiconductor substrate;
a plurality of electrode sets each of which consists of first, second, third and fourth transfer electrodes, each of said transfer electrodes being formed on an insulating layer formed on a top surface of said top layer, said first, second, third and fourth transfer electrodes being regularly arranged to form a four-phase vertical CCD shift register, each of said first electrodes comprising a lower portion formed on said insulating layer, a first upper portion formed on a neighboring one of said fourth electrodes through an insulating film and a second upper portion formed on a neighboring one of said second electrodes through an insulating film, each of said third electrodes comprising a lower portion formed on said insulating layer, a first upper portion formed on a neighboring one of said second electrodes through an insulating film and a second upper portion formed on a neighboring one of said fourth electrodes; and
a means for making a potential well under each of said first transfer electrodes deeper than a next potential well under a neighboring one of said second transfer electrodes.
A CCD imager according to Claim 7 wherein said means comprises a plurality of first conductivity type regions each of which is formed in said top layer of said first conductivity type under a unique one of said first transfer electrodes, an impurity concentration of said first conductivity type regions being higher than an impurity concentration of said top layer.
A CCD imager according to Claim 8 wherein said imager further comprises an array of photosensitive imaging elements connected with said vertical shift registers, and a means for supplying a first drive pulse signal to said first transfer electrodes, a second drive pulse signal to said second transfer electrodes, a third drive pulse signal to said third transfer electrodes, and a fourth drive pulse signal to said fourth transfer electrodes.